1 *> \brief \b CLATRD reduces the first nb rows and columns of a symmetric/Hermitian matrix A to real tridiagonal form by an unitary similarity transformation.
3 * =========== DOCUMENTATION ===========
5 * Online html documentation available at
6 * http://www.netlib.org/lapack/explore-html/
9 *> Download CLATRD + dependencies
10 *> <a href="http://www.netlib.org/cgi-bin/netlibfiles.tgz?format=tgz&filename=/lapack/lapack_routine/clatrd.f">
12 *> <a href="http://www.netlib.org/cgi-bin/netlibfiles.zip?format=zip&filename=/lapack/lapack_routine/clatrd.f">
14 *> <a href="http://www.netlib.org/cgi-bin/netlibfiles.txt?format=txt&filename=/lapack/lapack_routine/clatrd.f">
21 * SUBROUTINE CLATRD( UPLO, N, NB, A, LDA, E, TAU, W, LDW )
23 * .. Scalar Arguments ..
25 * INTEGER LDA, LDW, N, NB
27 * .. Array Arguments ..
29 * COMPLEX A( LDA, * ), TAU( * ), W( LDW, * )
38 *> CLATRD reduces NB rows and columns of a complex Hermitian matrix A to
39 *> Hermitian tridiagonal form by a unitary similarity
40 *> transformation Q**H * A * Q, and returns the matrices V and W which are
41 *> needed to apply the transformation to the unreduced part of A.
43 *> If UPLO = 'U', CLATRD reduces the last NB rows and columns of a
44 *> matrix, of which the upper triangle is supplied;
45 *> if UPLO = 'L', CLATRD reduces the first NB rows and columns of a
46 *> matrix, of which the lower triangle is supplied.
48 *> This is an auxiliary routine called by CHETRD.
56 *> UPLO is CHARACTER*1
57 *> Specifies whether the upper or lower triangular part of the
58 *> Hermitian matrix A is stored:
59 *> = 'U': Upper triangular
60 *> = 'L': Lower triangular
66 *> The order of the matrix A.
72 *> The number of rows and columns to be reduced.
77 *> A is COMPLEX array, dimension (LDA,N)
78 *> On entry, the Hermitian matrix A. If UPLO = 'U', the leading
79 *> n-by-n upper triangular part of A contains the upper
80 *> triangular part of the matrix A, and the strictly lower
81 *> triangular part of A is not referenced. If UPLO = 'L', the
82 *> leading n-by-n lower triangular part of A contains the lower
83 *> triangular part of the matrix A, and the strictly upper
84 *> triangular part of A is not referenced.
86 *> if UPLO = 'U', the last NB columns have been reduced to
87 *> tridiagonal form, with the diagonal elements overwriting
88 *> the diagonal elements of A; the elements above the diagonal
89 *> with the array TAU, represent the unitary matrix Q as a
90 *> product of elementary reflectors;
91 *> if UPLO = 'L', the first NB columns have been reduced to
92 *> tridiagonal form, with the diagonal elements overwriting
93 *> the diagonal elements of A; the elements below the diagonal
94 *> with the array TAU, represent the unitary matrix Q as a
95 *> product of elementary reflectors.
96 *> See Further Details.
102 *> The leading dimension of the array A. LDA >= max(1,N).
107 *> E is REAL array, dimension (N-1)
108 *> If UPLO = 'U', E(n-nb:n-1) contains the superdiagonal
109 *> elements of the last NB columns of the reduced matrix;
110 *> if UPLO = 'L', E(1:nb) contains the subdiagonal elements of
111 *> the first NB columns of the reduced matrix.
116 *> TAU is COMPLEX array, dimension (N-1)
117 *> The scalar factors of the elementary reflectors, stored in
118 *> TAU(n-nb:n-1) if UPLO = 'U', and in TAU(1:nb) if UPLO = 'L'.
119 *> See Further Details.
124 *> W is COMPLEX array, dimension (LDW,NB)
125 *> The n-by-nb matrix W required to update the unreduced part
132 *> The leading dimension of the array W. LDW >= max(1,N).
138 *> \author Univ. of Tennessee
139 *> \author Univ. of California Berkeley
140 *> \author Univ. of Colorado Denver
143 *> \date September 2012
145 *> \ingroup complexOTHERauxiliary
147 *> \par Further Details:
148 * =====================
152 *> If UPLO = 'U', the matrix Q is represented as a product of elementary
155 *> Q = H(n) H(n-1) . . . H(n-nb+1).
157 *> Each H(i) has the form
159 *> H(i) = I - tau * v * v**H
161 *> where tau is a complex scalar, and v is a complex vector with
162 *> v(i:n) = 0 and v(i-1) = 1; v(1:i-1) is stored on exit in A(1:i-1,i),
163 *> and tau in TAU(i-1).
165 *> If UPLO = 'L', the matrix Q is represented as a product of elementary
168 *> Q = H(1) H(2) . . . H(nb).
170 *> Each H(i) has the form
172 *> H(i) = I - tau * v * v**H
174 *> where tau is a complex scalar, and v is a complex vector with
175 *> v(1:i) = 0 and v(i+1) = 1; v(i+1:n) is stored on exit in A(i+1:n,i),
176 *> and tau in TAU(i).
178 *> The elements of the vectors v together form the n-by-nb matrix V
179 *> which is needed, with W, to apply the transformation to the unreduced
180 *> part of the matrix, using a Hermitian rank-2k update of the form:
181 *> A := A - V*W**H - W*V**H.
183 *> The contents of A on exit are illustrated by the following examples
184 *> with n = 5 and nb = 2:
186 *> if UPLO = 'U': if UPLO = 'L':
188 *> ( a a a v4 v5 ) ( d )
189 *> ( a a v4 v5 ) ( 1 d )
190 *> ( a 1 v5 ) ( v1 1 a )
191 *> ( d 1 ) ( v1 v2 a a )
192 *> ( d ) ( v1 v2 a a a )
194 *> where d denotes a diagonal element of the reduced matrix, a denotes
195 *> an element of the original matrix that is unchanged, and vi denotes
196 *> an element of the vector defining H(i).
199 * =====================================================================
200 SUBROUTINE CLATRD( UPLO, N, NB, A, LDA, E, TAU, W, LDW )
202 * -- LAPACK auxiliary routine (version 3.4.2) --
203 * -- LAPACK is a software package provided by Univ. of Tennessee, --
204 * -- Univ. of California Berkeley, Univ. of Colorado Denver and NAG Ltd..--
207 * .. Scalar Arguments ..
209 INTEGER LDA, LDW, N, NB
211 * .. Array Arguments ..
213 COMPLEX A( LDA, * ), TAU( * ), W( LDW, * )
216 * =====================================================================
219 COMPLEX ZERO, ONE, HALF
220 PARAMETER ( ZERO = ( 0.0E+0, 0.0E+0 ),
221 $ ONE = ( 1.0E+0, 0.0E+0 ),
222 $ HALF = ( 0.5E+0, 0.0E+0 ) )
224 * .. Local Scalars ..
228 * .. External Subroutines ..
229 EXTERNAL CAXPY, CGEMV, CHEMV, CLACGV, CLARFG, CSCAL
231 * .. External Functions ..
234 EXTERNAL LSAME, CDOTC
236 * .. Intrinsic Functions ..
239 * .. Executable Statements ..
241 * Quick return if possible
246 IF( LSAME( UPLO, 'U' ) ) THEN
248 * Reduce last NB columns of upper triangle
250 DO 10 I = N, N - NB + 1, -1
256 A( I, I ) = REAL( A( I, I ) )
257 CALL CLACGV( N-I, W( I, IW+1 ), LDW )
258 CALL CGEMV( 'No transpose', I, N-I, -ONE, A( 1, I+1 ),
259 $ LDA, W( I, IW+1 ), LDW, ONE, A( 1, I ), 1 )
260 CALL CLACGV( N-I, W( I, IW+1 ), LDW )
261 CALL CLACGV( N-I, A( I, I+1 ), LDA )
262 CALL CGEMV( 'No transpose', I, N-I, -ONE, W( 1, IW+1 ),
263 $ LDW, A( I, I+1 ), LDA, ONE, A( 1, I ), 1 )
264 CALL CLACGV( N-I, A( I, I+1 ), LDA )
265 A( I, I ) = REAL( A( I, I ) )
269 * Generate elementary reflector H(i) to annihilate
273 CALL CLARFG( I-1, ALPHA, A( 1, I ), 1, TAU( I-1 ) )
279 CALL CHEMV( 'Upper', I-1, ONE, A, LDA, A( 1, I ), 1,
280 $ ZERO, W( 1, IW ), 1 )
282 CALL CGEMV( 'Conjugate transpose', I-1, N-I, ONE,
283 $ W( 1, IW+1 ), LDW, A( 1, I ), 1, ZERO,
285 CALL CGEMV( 'No transpose', I-1, N-I, -ONE,
286 $ A( 1, I+1 ), LDA, W( I+1, IW ), 1, ONE,
288 CALL CGEMV( 'Conjugate transpose', I-1, N-I, ONE,
289 $ A( 1, I+1 ), LDA, A( 1, I ), 1, ZERO,
291 CALL CGEMV( 'No transpose', I-1, N-I, -ONE,
292 $ W( 1, IW+1 ), LDW, W( I+1, IW ), 1, ONE,
295 CALL CSCAL( I-1, TAU( I-1 ), W( 1, IW ), 1 )
296 ALPHA = -HALF*TAU( I-1 )*CDOTC( I-1, W( 1, IW ), 1,
298 CALL CAXPY( I-1, ALPHA, A( 1, I ), 1, W( 1, IW ), 1 )
304 * Reduce first NB columns of lower triangle
310 A( I, I ) = REAL( A( I, I ) )
311 CALL CLACGV( I-1, W( I, 1 ), LDW )
312 CALL CGEMV( 'No transpose', N-I+1, I-1, -ONE, A( I, 1 ),
313 $ LDA, W( I, 1 ), LDW, ONE, A( I, I ), 1 )
314 CALL CLACGV( I-1, W( I, 1 ), LDW )
315 CALL CLACGV( I-1, A( I, 1 ), LDA )
316 CALL CGEMV( 'No transpose', N-I+1, I-1, -ONE, W( I, 1 ),
317 $ LDW, A( I, 1 ), LDA, ONE, A( I, I ), 1 )
318 CALL CLACGV( I-1, A( I, 1 ), LDA )
319 A( I, I ) = REAL( A( I, I ) )
322 * Generate elementary reflector H(i) to annihilate
326 CALL CLARFG( N-I, ALPHA, A( MIN( I+2, N ), I ), 1,
333 CALL CHEMV( 'Lower', N-I, ONE, A( I+1, I+1 ), LDA,
334 $ A( I+1, I ), 1, ZERO, W( I+1, I ), 1 )
335 CALL CGEMV( 'Conjugate transpose', N-I, I-1, ONE,
336 $ W( I+1, 1 ), LDW, A( I+1, I ), 1, ZERO,
338 CALL CGEMV( 'No transpose', N-I, I-1, -ONE, A( I+1, 1 ),
339 $ LDA, W( 1, I ), 1, ONE, W( I+1, I ), 1 )
340 CALL CGEMV( 'Conjugate transpose', N-I, I-1, ONE,
341 $ A( I+1, 1 ), LDA, A( I+1, I ), 1, ZERO,
343 CALL CGEMV( 'No transpose', N-I, I-1, -ONE, W( I+1, 1 ),
344 $ LDW, W( 1, I ), 1, ONE, W( I+1, I ), 1 )
345 CALL CSCAL( N-I, TAU( I ), W( I+1, I ), 1 )
346 ALPHA = -HALF*TAU( I )*CDOTC( N-I, W( I+1, I ), 1,
348 CALL CAXPY( N-I, ALPHA, A( I+1, I ), 1, W( I+1, I ), 1 )